libvirt/docs/bugs.html.in
Daniel P. Berrangé 91205f1a86 docs: point users to gitlab for issue tracking
Currently we use the "Virtualization Tools" product in Red Hat Bugzilla
for issue tracking upstream. This changes to point people to GitLab for
issue tracking.

Note that Bugzilla still has plenty of bugs present against libvirt.
Triaging these to determine what is still valid will be a separate
exercise. Bugzilla will be locked to prevent creation of new issues
meanwhile.

Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2020-04-09 16:31:17 +01:00

162 lines
6.6 KiB
XML

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml">
<body>
<h1>Bug reporting</h1>
<ul id="toc"></ul>
<h2><a id="security">Security Issues</a></h2>
<p>
If you think that an issue with libvirt may have security
implications, <strong>please do not</strong> publicly
report it in the bug tracker, mailing lists, or irc. Libvirt
has <a href="securityprocess.html">a dedicated process for handling (potential) security issues</a>
that should be used instead. So if your issue has security
implications, ignore the rest of this page and follow the
<a href="securityprocess.html">security process</a> instead.
</p>
<h2><a id="bugtracking">Bug Tracking</a></h2>
<p>
If you are using libvirt binaries from a Linux distribution
check below for distribution specific bug reporting policies
first.
</p>
<h2><a id="general">General libvirt bug reports</a></h2>
<p>
Bugs in upstream libvirt code should be reported as issues in the
appropriate <a href="https://gitlab.com/libvirt">project on GitLab.</a>
Before submitting a ticket, check the existing tickets to see if
the bug/feature is already tracked.
</p>
<p>
It's always a good idea to file bug reports, as the process of
filing the report always makes it easier to describe the
problem, and the bug number provides a quick way of referring to
the problem. However, not everybody in the community pays frequent
attention to issues, so after you file a bug, asking questions
and submitting patches on <a href="contact.html">the libvirt
mailing lists</a> will increase your bug's visibility and
encourage people to think about your problem. Don't hesitate to
ask questions on the list, as others may know of existing
solutions or be interested in collaborating with you on finding
a solution. Patches are always appreciated, and it's likely
that someone else has the same problem you do!
</p>
<p>
If you decide to write code, though, before you begin please
read the <a href="hacking.html">contributor guidelines</a>,
especially the first point: "Discuss any large changes on the
mailing list first. Post patches early and listen to feedback."
Few development experiences are more discouraging than spending
a bunch of time writing a patch only to have someone point out a
better approach on list.
</p>
<ul>
<li><a href="https://gitlab.com/libvirt/libvirt/-/issues">View libvirt.git tickets</a></li>
<li><a href="https://gitlab.com/libvirt/libvirt/-/issues/new">New libvirt.git ticket</a></li>
</ul>
<p>
Note bugs in language bindings and other sub-projects should be
reported to their corresponding git repository rather than the
main libvirt.git linked above.
</p>
<h2><a id="distribution">Linux Distribution specific bug reports</a></h2>
<ul>
<li>
If you are using binaries from <strong>Fedora</strong>, enter
tickets against the <code>Fedora</code> product and
the <code>libvirt</code> component.
<ul>
<li><a href="http://bugzilla.redhat.com/buglist.cgi?component=libvirt&amp;product=Fedora">View Fedora libvirt tickets</a></li>
<li><a href="http://bugzilla.redhat.com/bugzilla/enter_bug.cgi?product=Fedora&amp;component=libvirt">New Fedora libvirt ticket</a></li>
</ul>
</li>
<li>
<p>
If you are using binaries from <strong>Red Hat Enterprise
Linux</strong>, enter tickets against the Red Hat Enterprise
Linux product that you're using (e.g., Red Hat Enterprise
Linux 6) and the <code>libvirt</code> component. Red Hat
bugzilla has <a href="http://bugzilla.redhat.com">additional guidance</a> about getting support if
you are a Red Hat customer.
</p>
</li>
<li>
<p>
If you are using binaries from another Linux distribution
first follow their own bug reporting guidelines.
</p>
</li>
<li>
<p>
Finally, if you are a contributor to another Linux
distribution and would like to have your procedure for
filing bugs mentioned here, please mail the libvirt
development list.
</p>
</li>
</ul>
<h2><a id="quality">How to file high quality bug reports</a></h2>
<p>
To increase the likelihood of your bug report being addressed it is
important to provide as much information as possible. When filing
libvirt bugs use this checklist to see if you are providing enough
information:
</p>
<ul>
<li>The version number of the libvirt build, or SHA1 of the GIT
commit</li>
<li>The hardware architecture being used</li>
<li>The name of the hypervisor (Xen, QEMU, KVM)</li>
<li>The XML config of the guest domain if relevant</li>
<li>For Xen hypervisor, the domain logfiles from /var/log/xen and
/var/log/libvirt/libxl</li>
<li>For QEMU/KVM, the domain logfile from /var/log/libvirt/qemu</li>
</ul>
<p>
If the bug leads to a tool linked to libvirt crash, then the best
is to provide a backtrace along with the scenario used to get the
crash, the simplest is to run the program under gdb, reproduce the
steps leading to the crash and then issue a gdb "bt -a" command to
get the stack trace, attach it to the bug. Note that for the
data to be really useful libvirt debug information must be present
for example by installing libvirt debuginfo package on Fedora or
Red Hat Enterprise Linux (with debuginfo-install libvirt) prior
to running gdb.</p>
<p>
It may also happen that the libvirt daemon itself crashes or gets stuck,
in the first case run it (as root) under gdb, and reproduce the sequence
leading to the crash, similarly to a normal program provide the
"bt" backtrace information to where gdb will have stopped.<br/>
But if libvirtd gets stuck, for example seems to stop processing
commands, try to attach to the faulty daemon and issue a gdb command
"thread apply all bt" to show all the threads backtraces, as in:</p>
<pre> # ps -o etime,pid `pgrep libvirt`
... note the process id from the output
# gdb /usr/sbin/libvirtd
.... some information about gdb and loading debug data
(gdb) attach $the_daemon_process_id
....
(gdb) thread apply all bt
.... information to attach to the bug
(gdb)
</pre>
</body>
</html>